public inbox for linux-xfs@vger.kernel.org
 help / color / mirror / Atom feed
From: Alex Elder <aelder@sgi.com>
To: Dave Chinner <david@fromorbit.com>
Cc: xfs@oss.sgi.com
Subject: Re: [PATCH 03/12] xfs: modify buffer item reference counting
Date: Wed, 19 May 2010 13:35:58 -0500	[thread overview]
Message-ID: <1274294158.2140.35.camel@doink> (raw)
In-Reply-To: <1274138668-1662-4-git-send-email-david@fromorbit.com>

On Tue, 2010-05-18 at 09:24 +1000, Dave Chinner wrote:
> From: Dave Chinner <dchinner@redhat.com>
> 
> The buffer log item reference counts used to take referenceѕ for every
> transaction, similar to the pin counting. This is symmetric (like the
> pin/unpin) with respect to transaction completion, but with dleayed logging
> becomes assymetric as the pinning becomes assymetric w.r.t. transaction
> completion.

Doing the reference counts this way is an improvement.

I have a question below.

. . .
> --- a/fs/xfs/xfs_buf_item.c
> +++ b/fs/xfs/xfs_buf_item.c
. . .
>  STATIC void
>  xfs_buf_item_unlock(
> @@ -514,73 +523,54 @@ xfs_buf_item_unlock(
>  
>  	bp = bip->bli_buf;
>  
> -	/*
> -	 * Clear the buffer's association with this transaction.
> -	 */
> +	/* Clear the buffer's association with this transaction. */
>  	XFS_BUF_SET_FSPRIVATE2(bp, NULL);
>  
>  	/*
> -	 * If this is a transaction abort, don't return early.
> -	 * Instead, allow the brelse to happen.
> -	 * Normally it would be done for stale (cancelled) buffers
> -	 * at unpin time, but we'll never go through the pin/unpin
> -	 * cycle if we abort inside commit.
> +	 * If this is a transaction abort, don't return early.  Instead, allow
> +	 * the brelse to happen.  Normally it would be done for stale
> +	 * (cancelled) buffers at unpin time, but we'll never go through the
> +	 * pin/unpin cycle if we abort inside commit.
>  	 */
>  	aborted = (bip->bli_item.li_flags & XFS_LI_ABORTED) != 0;
>  
>  	/*
> -	 * If the buf item is marked stale, then don't do anything.
> -	 * We'll unlock the buffer and free the buf item when the
> -	 * buffer is unpinned for the last time.
> +	 * Before possibly freeing the buf item, determine if we should
> +	 * release the buffer at the end of this routine.
> +	 */
> +	hold = bip->bli_flags & XFS_BLI_HOLD;
> +
> +	/* Clear the per transaction state. */
> +	bip->bli_flags &= ~(XFS_BLI_LOGGED | XFS_BLI_HOLD);
> +
> +	/*
> +	 * If the buf item is marked stale, then don't do anything.  We'll
> +	 * unlock the buffer and free the buf item when the buffer is unpinned
> +	 * for the last time.
>  	 */
>  	if (bip->bli_flags & XFS_BLI_STALE) {
> -		bip->bli_flags &= ~XFS_BLI_LOGGED;
>  		trace_xfs_buf_item_unlock_stale(bip);
>  		ASSERT(bip->bli_format.blf_flags & XFS_BLI_CANCEL);
> -		if (!aborted)
> +		if (!aborted) {
> +			atomic_dec(&bip->bli_refcount);

I notice that, unlike before, if you return via this path
the XFS_BLI_HOLD flag will have been turned off.  I guess
I don't know off hand whether this is an issue.  Can you
explain why it is not?

>  			return;
> +		}
>  	}
>  
. . .

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

  reply	other threads:[~2010-05-19 18:33 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-05-17 23:24 [PATCH 0/12] xfs: delayed logging V6 Dave Chinner
2010-05-17 23:24 ` [PATCH 01/12] xfs: Don't reuse the same transaction ID for duplicated transactions Dave Chinner
2010-05-17 23:24 ` [PATCH 02/12] xfs: allow log ticket allocation to take allocation flags Dave Chinner
2010-05-17 23:24 ` [PATCH 03/12] xfs: modify buffer item reference counting Dave Chinner
2010-05-19 18:35   ` Alex Elder [this message]
2010-05-19 22:37     ` Dave Chinner
2010-05-17 23:24 ` [PATCH 04/12] xfs: Clean up XFS_BLI_* flag namespace Dave Chinner
2010-05-19 19:09   ` Alex Elder
2010-05-17 23:24 ` [PATCH 05/12] xfs: clean up log ticket overrun debug output Dave Chinner
2010-05-19 19:16   ` Alex Elder
2010-05-19 22:41     ` Dave Chinner
2010-05-17 23:24 ` [PATCH 06/12] xfs: make the log ticket ID available outside the log infrastructure Dave Chinner
2010-05-17 23:24 ` [PATCH 07/12] xfs: Improve scalability of busy extent tracking Dave Chinner
2010-05-20 20:15   ` Alex Elder
2010-05-21  2:16     ` Dave Chinner
2010-05-21 20:59       ` Alex Elder
2010-05-17 23:24 ` [PATCH 08/12] xfs: Delayed logging design documentation Dave Chinner
2010-05-21 21:02   ` Alex Elder
2010-05-17 23:24 ` [PATCH 09/12] xfs: Introduce delayed logging core code Dave Chinner
2010-05-21 21:06   ` Alex Elder
2010-05-17 23:24 ` [PATCH 10/12] xfs: forced unmounts need to push the CIL Dave Chinner
2010-05-17 23:24 ` [PATCH 11/12] xfs: enable background pushing of " Dave Chinner
2010-05-17 23:24 ` [PATCH 12/12] xfs: Ensure inode allocation buffers are fully replayed Dave Chinner
2010-05-21 21:21 ` [PATCH 0/12] xfs: delayed logging V6 Alex Elder
2010-05-22  0:39   ` Dave Chinner
2010-05-24  0:30 ` Dave Chinner
  -- strict thread matches above, loose matches on Subject: below --
2010-05-07  5:40 [PATCH 0/12] xfs: delayed logging V5 Dave Chinner
2010-05-07  5:40 ` [PATCH 03/12] xfs: modify buffer item reference counting Dave Chinner

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1274294158.2140.35.camel@doink \
    --to=aelder@sgi.com \
    --cc=david@fromorbit.com \
    --cc=xfs@oss.sgi.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox